Struct SensorGraph 

pub struct SensorGraph {
    pub nodes: Vec<SensorNode>,
    pub edges: Vec<SensorEdge>,
    pub temporal_buffers: Vec<Vec<f32>>,
}

Graph-based sensor fusion substrate for distributed buoy arrays.

Fields

nodes: Vec<SensorNode>

Buoy positions and node features.

edges: Vec<SensorEdge>

Inter-buoy connections.

temporal_buffers: Vec<Vec<f32>>

Temporal feature buffer per node (ring of scalar samples; typically short-time acoustic energy or a spectral band).

Implementations

impl SensorGraph

pub fn new() -> Self

Create an empty sensor graph.

pub fn add_buoy(&mut self, node: SensorNode) -> usize

Add a buoy node. Returns its index.

pub fn connect(&mut self, from: usize, to: usize)

Connect two buoys; distance and delay derived from positions.

Uses Euclidean distance in (lat, lon, depth) space and a constant 1500 m/s sound speed. Production deployments should swap in geodesic

• SSP / BELLHOP-style propagation.

Panics

Panics if from or to are out of bounds.

pub fn connect_with( &mut self, from: usize, to: usize, distance_m: f64, propagation_delay_ms: f64, )

Connect with caller-supplied metrics (tests / offline graph builders).

pub fn connect_within_radius(&mut self, radius: f64)

Build a radius graph: undirected edges when distance ≤ radius.

pub fn connect_fully(&mut self)

Fully connected undirected graph (K-STEMIT-style small arrays).

pub fn push_temporal(&mut self, node_idx: usize, values: &[f32])

Append a feature snapshot to a node’s temporal buffer.

Panics

Panics if node_idx is out of bounds.

pub fn trim_temporal(&mut self, max_len: usize)

Cap each temporal buffer to the last max_len samples.

pub fn neighbor_indices(&self, node_idx: usize) -> Vec<usize>

Neighbor indices for a node (undirected).

pub fn neighbor_distances(&self, node_idx: usize) -> HashMap<usize, f64>

Direct-edge distance map from node_idx to each neighbor.

pub fn aggregate_neighbors(&self, node_idx: usize, hop: usize) -> Vec<f32>

GraphSAGE-style mean aggregation (unlearned baseline, hop-BFs).

Weighted by inverse distance on direct edges; multi-hop nodes without a direct edge use weight 1.0. Prefer crate::sensor_graph::GraphSageLayer for the learned path.

pub fn temporal_features(&self, node_idx: usize, window: usize) -> Vec<f32>

Last window temporal samples (left zero-padded).

pub fn fused_features(&self, node_idx: usize) -> Vec<f32>

Baseline fused feature: own + 1-hop mean-agg + last-16 temporal.

pub fn node_count(&self) -> usize

Number of nodes.

pub fn edge_count(&self) -> usize

Number of edges.

pub fn feature_dim(&self) -> usize

Feature dimension of node 0 (0 if empty). Assumes homogeneous features.

Trait Implementations

impl Clone for SensorGraph

fn clone(&self) -> Self

Returns a duplicate of the value.

fn clone_from(&mut self, source: &Self)

Performs copy-assignment from source.

impl Debug for SensorGraph

fn fmt(&self, f: &mut Formatter<'_>) -> Result

Formats the value using the given formatter.

impl Default for SensorGraph

fn default() -> Self

Returns the “default value” for a type.